Power Supply Reliability Targets Based on Improved Regression Prediction AlgorithmsPrediction Methods and Applications

This paper proposes an improved regression prediction algorithm that enhances the accuracy of predicting power supply reliability targets in distribution grids. First, the method constructs exponential and power regression models based on historical reliability indicator data. Then, it introduces a weighted fusion mechanism to optimize the combination of weights and construct the optimal improved regression prediction model. Compared to a single regression model, this method more effectively captures the nonlinear variation characteristics of power supply reliability indicators, thereby improving prediction accuracy. To validate the model's performance, the study selected a medium-voltage distribution network in an area of Shenzhen for empirical analysis. The results showed that the proposed model's predictions are highly consistent with the actual power supply reliability targets, demonstrating the method's accuracy and practical value.